Social acceptance of serious games for physical and cognitive training in older adults residing in ambient assisted living environments
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Journal of Public Health: From Theory to Practice
https://doi.org/10.1007/s10389-021-01524-y
ORIGINAL ARTICLE
Social acceptance of serious games for physical and cognitive
training in older adults residing in ambient assisted living
environments
Philipp Brauner 1 & Martina Ziefle 1
Received: 1 October 2020 / Accepted: 18 March 2021
# The Author(s) 2021
Abstract
Aims Although ambient assisted living (AAL) environments and serious games for healthcare have been proposed as solutions to
meet the changing demographics, the two approaches are rarely combined. We present the development and empirical evaluation
of two serious games for healthcare in AAL. The first uses a cooking scenario for training of cognitive functioning. The latter uses
a gardening scenario and motion capture for training agility and endurance. As the frequent lack of social acceptance is a major
challenge in consumer health technology, we integrated methods of technology acceptance research by means of the UTAUT2-
model and intention to use into the evaluation.
Subject and methods We developed both games utilizing user-centered and participatory design methods ranging from low-
fidelity paper prototypes to usability and acceptance evaluations of functional prototypes. In the final evaluation, each game
prototype was evaluated by 64 participants form different age groups.
Results The results show that although performance decreases with age, the performance attained in the games is not decisive for
social acceptance measured as intention to use. However, user diversity factors shape the evaluation of the games, and older
people and people with low technical competence are in danger of being excluded from using serious games for healthcare.
Conclusion Exercise games, if designed right, can mitigate the negative effects of demographic change. Nevertheless, user
diversity needs must be considered to ensure that the games are usable and used by a broad audience. The article concludes
with guidelines and open research questions for the design of serious games in AAL environments.
Keywords Seriousgames . Ambientassistedliving . Social acceptance . Technology acceptance . Older adults . Cognitivetraining
game . Exercise game
Introduction New consumer health informatics (CHI) technologies are a
solution to mitigate the negative consequences of the changing
The world’s changing demographics are a double-edged sword. demographics. Ambient assistant living (AAL) and serious
On the one hand, life expectancy is increasing through improve- games for healthcare are both CHI techniques aimed at mitigat-
ments in medicine, nutrition, and hygiene. On the other hand, the ing the negative effects of the demographic change. The former
viability of health and social systems is threatened by a declining aims to increase the safety, comfort, and quality of life of resi-
workforce, obesity, and sedentary behavior, as well as an increas- dents by integrating technology into their living environment and
ing number of beneficiaries who are older and thus more depen- to enable them to live autonomous and self-determined lives
dent on care (Ho et al. 1993; Oeppen and Vaupel 2002; despite chronic illnesses (Kleinberger et al. 2007). The latter
Giannakouris 2008; López-Otín et al. 2013). combines the motivational character of games with serious goals
to increase health awareness, stamina, body control, and memory
or to support rehabilitation measures.
* Philipp Brauner In this article we combine these two rarely connected
philipp.brauner@rwth-aachen.de branches and we present the development and empirical eval-
uation of two serious games for healthcare in a prototypical
1
Human-Computer Interaction Center, RWTH Aachen University, AAL residential environment: the first game addresses the
Campus-Boulevard 57, 52074 Aachen, Germany improvement of executive functioning through a cookingJ Public Health (Berl.): From Theory to Practice
game, and the second promotes physical activity in a garden- likelihood of diabetes, 60% lower likelihood of dementia,
themed exercise game. Furthermore, as social acceptance of 60% lower risk for stroke, and an increase in life expectancy
medical technology by users—especially in the case of tech- by 6 years. Physical exercise increases pulmonary function,
nology for end users—is a major challenge (Czaja et al. 2006; oxygen consumption, cardiovascular function, standing and
Flandorfer 2012), we apply methods from technology accep- walking balance, bone stability, and heart and other muscle
tance research to evaluate our game environments and to iden- strength, and improves the immune system (Aldwin et al.
tify individual factors that promote or limit the adoption of the 2006). In combination with a healthy diet, it reduces over-
game-based healthcare interventions in AAL. Key questions weight and obesity and thus associated diseases such as dia-
addressed in this article are the relationship between user fac- betes. Regular physical exercise reduces the likelihood or in-
tors and performance within both serious games, the influence tensity of cardiovascular diseases, and patients suffering from
of user factors and the attained performance on the evaluation chronic syndromes benefit from regular and lifelong exercise
of the games, how the game evaluation relates to the overall at home, such as a combination of aerobic activities and resis-
projected use, and the similarities and differences between the tance, flexibility, and balance exercises (Perez‐Terzic 2012).
two games under investigation. One hundred and fifty minutes of medium-intensity aerobic
The structure of this article is as follows: After this intro- exercise per week is sufficient for a significant positive impact
duction, we present an overview of the benefits of exercise, on health (Mendes et al. 2011), especially for children, older
serious games for healthcare as a method to stimulate exercise, adults, and people with overweight or obesity. Although the
ambient assisted living, and technology acceptance research. frequency, intensity, and duration of exercise can be opti-
Next, we present the two game prototypes developed for this mized, the key is that some physical activity is always better
work. After that we illustrate our experimental framework for than none (Saint-Maurice et al. 2018).
the evaluation, the samples, and the key findings. Finally, we Physical exercise also has benefits that are less apparent. It
discuss our work, its implications, and its limitations and ideas reduces the probability of silent (unnoticed) brain infarcts in
for future work. the elderly by 40% (Willey et al. 2011) and can be as effective
against the occurrence of migraines as an anti-migraine drug
(Varkey et al. 2011). A more active lifestyle with social inter-
Related work action, listening to music, and physical activity lowers ten-
sion, improves mood, and increases perceived energy levels
As this article combines concepts and methods form multiple (Thayer et al. 1994). Exercise can also reduce the symptoms
distinct domains, this section gives a broad overview and ar- of depression, although antidepressants and psychotherapy
gues the benefits of physical and cognitive exercise, and pre- are found to be most effective (Cooney et al. 2013).
sents related work on serious games for healthcare, AAL, and Physical exercise also has short-term effects on cognition.
finally technology acceptance research. Working memory performance is directly increased after
30 minutes of exercise (McMorris et al. 2011). Regular aero-
Benefits of physical and cognitive exercise bic exercise improves executive functioning for healthy peo-
ple, and older adults show improved task switching ability,
Aging is “characterized by a progressive loss of physiological higher selective attention, higher working memory capacity,
integrity, leading to impaired function and increased vulner- and better overall performance (Guiney and Machado 2013).
ability to death” (López-Otín et al. 2013, p. 1194), and affects Remarkably, physical exercise in midlife reduces the risk of
all aspects of human life. It is associated with an increased cognitive decline, mild cognitive impairment, and dementia
likelihood of chronic illnesses or disabilities such as conges- later in life (Ahlskog et al. 2011).
tive heart failures (Ho et al. 1993), stroke, diabetes mellitus For the domain of cognitive training, research on the effec-
(Hader et al. 2004), mild cognitive impairment (MCI), demen- tiveness and medical implications is less conclusive. On the
tia, or Alzheimer’s disease (AD) (Gao et al. 1998). Besides one hand, computerized cognitive training (CCT) programs
age-related diseases, overweight and obesity are increasing are postulated to be suitable, effective, and necessary for
risk factors and are threatening life expectancy (Mann 2005). healthy brain aging by stimulating neuroplasticity (Ball et al.
Physical inactivity is associated with several diseases in- 2004; Heyn et al. 2004) and may contribute to healthy brain
cluding cancer, diabetes, hypertension, and coronary or cere- aging (Shah et al. 2017).
brovascular diseases, and reduces life expectancy (Knight
2012). However, health and life expectancy can be increased Serious games and serious games for healthcare
by maintaining a healthy body weight, a healthy diet, not
smoking, low alcohol intake, and regular exercise, as shown The main idea behind serious games (for healthcare and for
by a 35-year long-term study (Elwood et al. 2013). People other domains) is built on the Premack principle (Premack
who showed all or almost all habits had a 73% lower 1959): the likelihood of performing an unpleasant and thusJ Public Health (Berl.): From Theory to Practice
infrequent activity—such as training and exercise—can be technology enthusiasts, the widespread penetration of AAL
increased by linking it with a more pleasant and thus more is still a bold vision of the future, and many fundamental
frequently performed activity—such as playing games. It in- questions regarding acceptance, perception, use, and interac-
volves packaging less liked and therefore less practiced activ- tion styles are still open (Cook and Das 2007; Lindley et al.
ities with games to increase the motivation and thus the like- 2008; Cook et al. 2009).
lihood of the activity being carried out. Serious games can be To study how AAL environments are perceived by future
defined as “a game in which education (in its various forms) is inhabitants and how these environments need to be designed,
the primary goal, rather than entertainment” (Michael and and to evaluate how these will shape future living, researchers
Chen 2006). In contrast, the related concept of gamification construct prototypical living labs to simulate the envisioned
(Deterding et al. 2011) does not use games as packaging models. These labs are built to test functional and nonfunc-
around activities, but certain real-world activities are incentiv- tional properties of innovative technologies and evaluation by
ized by game-like elements (e.g., points, badges). potential users. Potential residents thus not only have the op-
Examples of serious games for healthcare are numerous portunity to be part of the evaluation of the technology, but
and address various medical domains such as mental health also can play an active role by contributing feedback, fears,
(Fleming et al. 2017), rehabilitation (Nap and Diaz-Orueta and ideas regarding the modern technology, and the residents
2015; Bonnechère et al. 2016), physical exercise for the elder- can thereby shape their future technology-augmented living
ly (Konstantinidis et al. 2016), and even training of medical environments. We will present the test environment used for
professionals (Gentry et al. 2019). this work in more detail in the Method section.
Also, there are already many available commercial game
titles for physical exercise or cognitive training (e.g., a variety Which factors influence the use of technology?
of mobile games and console titles such as Wii Fit and Brain
Boost for Nintendo, Body and Brain Connection for Microsoft The main goal of technology acceptance research is predicting
Xbox, or Smart As… for Sony PlayStation). There is evidence whether and how a given technology will be adopted by users
that exercise gaming at home has positive effects on health, and which technological (capabilities, design, interfaces, etc.)
pain perception, and overall well-being (Warburton et al. and user factors (age, gender, attitude towards technology) are
2007; Whitehead et al. 2010; Franco et al. 2012; Brauner relevant to the actual use (Davis 1993; Rogers 2003).
and Ziefle 2020), but the effectiveness and transferability of A cornerstone of this research domain is Davis’ technology
cognitive training games is not as well established. A meta- acceptance model (TAM) aimed at predicting the actual use of
analysis found small but significant positive effects on pro- a technology through the intention to use that can be measured
cessing speed, working memory, and verbal and nonverbal in advance (Davis 1993). TAM has revealed a strong relation-
memory, but no effect on attention or executive function ship between the perceived usefulness and ease of use of the
(Lampit et al. 2014). Also, home-based and too frequent train- technology, and the projected and the later actual use.
ing was found to be ineffective. A subsequent meta-analysis Over time, numerous other models have emerged and have
found that commercial video games increased cognitive per- integrated additional explanatory factors, increased the predictive
formance, such as attention and problem-solving skills. power, and opened new areas of application. However, this has
However, the performance gains were limited to the specific led to a variety of highly specific models that are rarely transfer-
task in the games and had no further positive effects (Choi able to other areas (such as TAM2, TAM3, and unified theory of
et al. 2020). Consequently, more research on the effectiveness acceptance and use of technology [UTAUT]) (Venkatesh et al.
of CCT is needed. 2003). Regarding our work, there is no established model for
studying serious games for healthcare or for evaluating AAL
Ambient assisted living environments. Nevertheless, the existing and established models
provide at least some guidance as to which factors are suitable for
Apart from serious games for healthcare, AAL environments studying the social acceptance of serious games for healthcare in
are another active research area aimed at mitigating the nega- AAL environments. In this work, we use the UTAUT2 model
tive effects of the changing demographics: by augmenting the (Venkatesh et al. 2012), as it is the first model that integrates the
living environments with smart sensors and actuators, elderly hedonic evaluation of a technology. In this model, later use is
(or chronically ill) people can stay independent in their own predicted through the intention to use, which is explained by the
living environments for longer (Raisinghani et al. 2006; seven factors of performance expectancy (perceived benefits of
Kleinberger et al. 2007; Rashidi and Mihailidis 2013). These the technology), effort expectancy (how complicated it is to use),
environments reduce the costs of healthcare systems while facilitating conditions (is the technology compatible with my
increasing individuals’ independent living and self-determina- environment, can others help me in case of trouble), social
tion. However, despite its immense potential and the increas- influence (how do my peers rate my use of the technology),
ing use of smart home components in the homes of a few price–value trade-off (is the technology worth the money),J Public Health (Berl.): From Theory to Practice
hedonic motivation (how much fun is the technology to use), and entertaining, and rewarding. Personalized messages may in-
habit (can I integrate the use into my daily routine). crease engagement with the game as well (“Linus, your score
Social acceptance is both shaped by the evaluation of the is not so good; keep on trying.”).
system and mediated by individual user factors. Numerous More recently, Ijsselsteijn et al. (2007) identified require-
studies have shown that age is associated not only with less ments, benefits, and opportunities for game design for elderly
ease of use in interacting with novel technologies, but also users. For example, they suggest that the boundaries of the
with lower perceived usefulness and overall acceptance visual game environment should adjust to the individual visu-
(Selwyn et al. 2003; Arning and Ziefle 2007; Schreder et al. al decrement. Also, information and feedback should use mul-
2013). Furthermore, women, on average, report lower interest tiple modalities (e.g., visual and auditory feedback) to account
and skill in dealing with technical systems. This is especially for preferences and individual limitations. Memory load and
relevant for the design of technology-based interventions in cognitive requirements should be kept to a minimum, for ex-
the healthcare sector, as these interventions should be usable ample, by avoiding the need to remember information from
and useful for all potential users, regardless of age, gender, or various parts of the game). Lastly, players should be given
(perceived) technical self-efficacy (Wilkowska et al. 2018). time to get used to the game and learn the basic interactions.
During this time in particular, encouraging feedback should be
Key challenges provided to support insecure players.
The first guidelines for exercise games that use body sensors
Even though the two areas of AAL and serious games for were presented by Whitehead et al. (2010). Games should detect
healthcare are established concepts in research that are in- whether body movements were correctly executed, and short-
creasingly flowing into the market, the intersection has re- ened movements should be rejected. Activities and gestures
ceived little attention. Consequently, our research objective should be designed to use larger muscles of the body, such as
is to develop healthcare games in technology-augmented en- the legs. The games should support players gaining experience
vironments, to evaluate how elderly people—who are often and playing the games for longer periods, for example by pro-
reluctant to deal with innovative technology—interact with viding incentives, as experience and frequent use yields higher
these games, and to identify the predictors for social accep- physical benefits. These incentives should have a lasting effect,
tance and adoption by means of technology acceptance re- not just for hours or days, but for months and years. Ideally, these
search. Thus, this work is at the intersection of technology incentives should not be related to the health benefits but should
development for AAL, healthcare, and user research. Here, not be an abstraction. Furthermore, there must be regular recu-
we combine the knowledge about the benefits of exercise, peration periods within the games to allow players to rest be-
serious games as a method to promote activities usually per- tween active phases.
ceived as unpleasant, the invisible integration of interfaces in Gerling et al. (2012) presented guidelines specifically for
AAL environments, and user and acceptance modeling to un- full-body motion exercise games for the elderly. For example,
derstand whether this approach increases the likelihood of interactions within the game should take account of potential
exercising and which factors are the largest contributors to individual physical or cognitive impairments, games should
projected acceptance and thus later use. adapt to limitations in players’ range of motion, and overex-
Designing healthcare games for the elderly is difficult, as ertion should be managed by alternating between physically
many distinct aspects must be considered. Besides the identi- intense and less challenging parts. Further, the difficulty and
fication of suitable training domains, the games must be fun challenge of the game should not be fixed but should adapt to
and entertaining, adapt to the various age-related changes such players’ abilities. Movement gestures should be easy to learn
as lower information processing speed and reaction times, and easy to remember, and ideally there should be a natural
altered vision (Fisk et al. 2009), and different interests and mapping between gestures and real-world activities. Players
often less ability to interact with computers compared to youn- should be guided by initial tutorials and later hints at required
ger adults. actions. Finally, games should support independent play,
To facilitate these numerous, often conflicting, require- meaning that routines to start the game should be easy and
ments, guidelines and recommendations have evolved over intuitive, and no assistance should be needed.
the years. One of the earliest guidelines for serious games
for healthcare emerged in the early 1980s when Weisman
(1983) modified and evaluated games on the Apple II for Method
institutionalized elderly.
Games should be slowed down and be focused on single Our work focuses on the concept of prevention before treatment:
tasks; on-screen objects should be well-defined and larger, the games presented here to do not address a specific disease, but
and additional hints should be provided to support interaction. rather overall physical or cognitive fitness. We developed two
Playing with others has also been found to be fun, different game prototypes to study the communalities andJ Public Health (Berl.): From Theory to Practice
differences between distinct kinds of serious games for healthcare For both game prototypes we applied a strictly iterative,
in AAL environments. The first game addresses general physical user-centered, and participatory design approach. Starting
fitness, agility, endurance, and body control (“Fitness Farm”), with low-fidelity paper prototypes (Snyder 2003), we evalu-
whereas the second game is designed to promote executive func- ated the game concept and the basic interactions with potential
tioning (“Cook it Right”). Both games incorporate the sensors and future users of the games. Using interviews, we captured feed-
display surfaces embedded in the AAL lab presented in the fol- back and ideas for improving the game design and the inter-
lowing, and thus require no further interaction devices and no actions. In subsequent iterations, we increased the maturity of
additional hardware setup. the game prototypes and eventually arrived at fully functional
game prototypes that were then evaluated in the final evalua-
Ambient assisted living lab tions presented below. Key improvements were, for example,
implicit feedback on the remaining playing time through a
We developed and evaluated the serious games in RWTH changing daylight in the case of the motion game, and a focus
Aachen University’s AAL lab. It replicates a typical 25 m2 on turn-based interactions compared to a timed user interface
living room with windows, a parquet floor, two couches and a in the cognitive training game. An in-depth presentation of the
table, ceiling lamps, and pictures and shelves.1 However, this design process and related findings for each game can be
lab has integrated various novel input and output technologies found in Wittland et al. (2015) and Brauner and Ziefle
to make the room more supportive. It assists future residents in (2020). Figure 1 shows a picture of an early (left) and the final
monitoring their diseases by providing novel communication prototype (right) for the motion (top) and cognitive (bottom)
channels to family, friends, and caregivers, or simply by in- training game.
creased comfort. Sensors invisibly embedded in the parquet
floor detect potential falls and may alert family members or Experimental procedure
caregivers (Klack et al. 2011a; Leusmann et al. 2011), a hid-
den invisible infrared camera measures a person’s body tem- We applied a congruent research design to evaluate both game
perature for detecting illnesses (Klack et al. 2011b), and a prototypes from the perspective of potential future users. Sixty-
large multi-touch wall can visually extend to room (Kasugai four participants evaluated each prototype in a between-subject
et al. 2010), support novel interaction techniques, and may design. They were recruited through personal social networks
provide telemedicine consultation services (Beul et al. 2012). and public notices in the city. To determine whether age or gender
What is essential about this environment is that the moni- has a particular influence on performance or social acceptance, we
toring technology is integrated invisibly, and the devices do recruited both younger and older adults, as well as male and female
not have to be set up by the users. subjects. For the experiment, we first introduced the participants to
the AAL lab, its core features, and how they could now contribute
to the design of serious games in this environment. Second, using a
Game concept and implementation
questionnaire, we surveyed the participants’ characteristics includ-
ing age, gender, chronic illnesses, technical self-efficacy, and in-
In the first game, the player is in a garden environment and the
clination to play. Next, we introduced the games, and the partici-
task is to collect several types of fruits and vegetables through
pants played three game rounds while performance metrics were
full-body motion gestures. Microsoft Kinect sensors hidden in
captured using log files. Third, we measured the perception of the
the AAL lab capture the player’s body postures. These ges-
game on a second and final questionnaire. We concluded the
tures were developed in cooperation with a physiotherapist
experimental session with a brief and informal chat about the game
and address harmless and often insufficiently trained move-
and the living lab. Participation was voluntary and—aside from
ments such as stretching or bending. Throughout the game’s
cookies—not rewarded. Figure 2 illustrates the experimental pro-
levels, different and increasingly difficult body gestures are
cedure and the captured variables.
introduced and practiced.
In the second game, players take on the role of a cook.
User factors
They must first memorize recipes and then recook them in a
graphical, touch-based interface. The difficulty of the recipes
On a questionnaire before the game intervention, we captured
increases with the levels of the game, for example, by requir-
the participants’ age and their gender, as both factors are often
ing more ingredients or more branched work steps. The game
linked with different usage, perceived ease of use, self-effica-
uses the multi-touch display surfaces of the AAL lab, namely
cy, and performance when interacting with information and
a conventional tablet device, a sofa table with an integrated
communication technology (Arning and Ziefle 2007;
display, or the large wall-sized display.
Schreder et al. 2013). We constructed a sample with both
1
A video tour of the AAL lab is available online: https://vimeo.com/ younger and older, and male and female participants (see de-
31951636. scription of the sample below).J Public Health (Berl.): From Theory to Practice
Fig. 1 Illustration of the iterative Final, functional prototype
development process of both
Fitness Farm (game one)
games
Motion-based interaction
Physical exercise game
Cook it Right (game two)
Touch based interactions
Cognitive training game
Iterative, user-centered, participatory design process
We also measured additional explanatory variables (not Need for achievement (NfA) This construct describes the gen-
used for constructing the sample) that prior work had identi- eral and stable tendency to work on relevant tasks with energy
fied as relevant predictors for performance or acceptance. and perseverance until successful completion, and relates to
the choice of tasks and the performance attained in the tasks
Chronic illnesses As people with chronic illness may have (Schuler and Prochaska 2001; Elliot and Dweck 2005). We
lower performance levels (e.g., due to physical limitations) suspected that a lower NfA could lead to lower performance in
or higher evaluation (e.g., higher perceived benefits), we the games or lower acceptance of games because of the com-
asked whether the participants suffered from chronic diseases, petitive components they contain. We used a short form with
and if so, which ones. four items of Schuler’s scale, for example, “I am attracted to
difficult problems.” The internal reliability of the scale was
Self-efficacy in interacting with technology (SET) Self-efficacy very good (α ≥ .829).
is the domain-specific belief, based on the perception of one’s own
competence, that one is successful in a certain activity. Women Gaming frequency (GF) Obviously, the general tendency to
and older adults often report lower scores in SET, which results in play will also affect the performance in and evaluation of
less motivation to deal with or lower performance in interacting serious games. To determine this, we asked the participants
with technical devices. We used the four items on a 6-point Likert for the current game frequency across multiple game domains
scale from Beier’s SET scale with the highest item-total correla- and media, such as games of skill and board and card games,
tion. The internal reliability was very good (α ≥ .830). but also games mediated through game consoles and cell
Fig. 2 Experimental procedure User Factors:
Pre survey
for evaluating the two serious - Age - Self-e cacy in interacting
- Gender with technology
games for healthcare in AAL - Chronic illness - Need for
environments - Gaming frequency Achievement
Game intervention (between-subject design)
Game Experience
Training of GAME 1 Training of GAME 2
physical skills cognitive skills Perf. Level 1
Perf. Level 2
64 participants 64 participants
17-84 years 16-84 years Perf. Level 3
32 male, 32 female 34 male, 30 female Increasing complexity
Post survey
Evaluation of the game (UTAUT2):
- Performance Expectancy Average
- E ort Expectancy - Hedonic Motivation performance
- Facilitating Conditions - Price-Value Tradeoff
- Habit - Intention To UseJ Public Health (Berl.): From Theory to Practice
phones, as well as ball and outdoor games. The scale consists Description of the samples
of nine items such as “I frequently play board games.” It had
very good internal reliability (α ≥ .730) and was further relat- The sample for the physical exercise game included 64 partic-
ed to items such as “I enjoy playing games” in prior work ipants (32 male, 32 female) aged 17–85 years (M = 43.2,
(Brauner et al. 2015). SD = 19.6), with no correlation between age and gender
All scales were measured on 6-point Likert items ranging (r = .110, p = .386 > .05). In our sample, age was associated
from 0 to 5 (max). with a decline in reported SET (r = −.548, p < .001) and lower
Dependent variables GF (r = −.569, p < .001). In addition, chronic diseases in-
creased with age (ρ = .406, p < .001), and the most frequently
For the evaluation, we measured (a) the participants’ mentioned diseases were asthma, hypertension, and diabetes.
performance in the games using logfiles, (b) the usability, per- Further, a link was observed between gender and SET (ρ =
ception, and social acceptance of the games using the UTAUT2 −.265, p = .035 < .001), with women reporting lower levels of
model in post-game surveys, and (c) qualitative observations self-efficacy than men. GF, however, was unrelated to gender
during the participants’ play and informal talks afterwards. in this sample (ρ = −.160, p = .206). NfA was not related to
The performance metrics differed between games. In either age (r = −.055, p = .672 > .05) or gender (ρ = −.139,
Fitness Farm each level lasted 90 s and we counted the num- p = .282 > .05). Table 1 (left) presents the correlations within
ber of collected objects as performance metric (higher number the first sample.
means higher performance). In Cook It Right we measured the The sample of the cognitive training game also consisted of
time the participants needed to successfully finish the recipes 64 participants (34 male, 30 female) aged 16–84 years (M =
(shorter duration means higher performance). 41.6, SD = 18.4). Again, there was no correlation between age
For the evaluation of both games, we designed a survey and gender (ρ = −.065, p = .611 > .05), and age was linked to
based on the UTAUT2 model (see above). First, we measured lower SET (r = −.490, p < .001) and lower GF (r = .722,
the evaluation of the game on the seven dimensions perfor- p < .001). Contrary to the first sample, NfA was linked to
mance expectancy, effort expectancy, facilitating conditions, age (r = −.327, p < .001) and gender (ρ = −.258, p = .042
social influence, price–value trade-off, hedonic motivation, < .05), with older people and women reporting being less per-
and habit each on three 6-point Likert items. Next, we asked formance-motivated. Again, gender was related to SET, with
for the intention to use the game as an indicator of social women reporting lower scores than men (ρ = −.387, p = .002
acceptance using three 3-point Likert items (“I would like to < .05). However, gender was not related to GF (ρ = −.219,
play the game again”). p = .082 > .05). Table 1 shows an overview of the correlations.
As this work focuses on the social acceptance of serious Next, the participants’ performance and their evaluation of
games in AAL, we did not measure health-related outcomes of the game is analyzed.
the games. Yet, based on the presented related work, we argue
that if the games are used to exercise, they will also reveal
positive effects on health and well-being.2 Results
Statistical analysis The structure of the results section is as follows: First, we
analyze the attained performance in both games (physical ex-
All subjective measures were captured on 6-point Likert ercise and cognitive training) and identify the strongest pre-
scales. For data analyses we used bivariate correlations dictors of performance. Second, we analyze how individual
(Pearson’s r, Spearman’s ρ), χ2, univariate and multivariate factors, including attained performance, relate to the overall
analyses of variance (ANOVA/MANOVA), and multiple lin- acceptance of the game.
ear regression for data analysis. We used and report Pillai’s
value V for omnibus tests of the MANOVAs. For the multiple Performance
linear regressions, we use the ENTER method, iteratively re-
move models with low standardized beta coefficients, and In this section, we analyze the players’ performance in the
exclude models with high variance inflation factors (VIF ≫ game, how it changed over the course of the game, and wheth-
1). We set the type I error rate (level of significance) to er user factors influenced performance.
α = .05 and report effect sizes as partial η2. Missing values For the motion-based exercise game, the number of collect-
are deleted list-wise on a per-test basis. ed objects was M = 22.8, SD = 6.5 (range 7–37) in the first
2 level (grabbing), M = 18.8, SD = 4.8 (9–29) in the second lev-
In fact, an analysis of the physical exercise game and its earlier prototype
found a mitigating effect on perceived pain in older adults shortly after playing el (grabbing + holding), and M = 17.0, SD = 4.9 (2–26) in the
the game (Brauner and Ziefle 2020). third level of the game (grabbing + holding + crossed over).J Public Health (Berl.): From Theory to Practice
Table 1 Characteristics of both samples. Left: motion-based physical exercise game. Right: touch-based cognitive training game
Fitness Farm – physical exercises Cook it Right – cognitive training
Desc. 1 2 3 4 5 Desc. 1 2 3 4 5
1 - Age 17–85 years – −.548** −.589** 17–86 years – −.490** −.722** −.347**
2 - Gender 32 m, 32w – −.265* 34 m, 30w – −.387* −.258*
3 - SET M=3.73 – .616** .393* M=3.64SD=1.29 – .499** .535**
SD=1.34
4 – GF M=3.73 – M=1.83SD=0.89 –
SD=1.34
5 - NfA M=3.73 – M=3.42 –
SD=1.34 SD=1.15
The increasing difficulty through more complex targets gamers being faster, whereas higher NfA yielded higher per-
yielded fewer collected game objects. formance in the cognitive training game (r = .318).
In the cognitive training game, the average completion time However, the sample description (see Table 2) showed that
was M = 57.3 s, SD = 18.7 s (range 33.2–107.2 s) for the first most user factors were interdependent, so it is not clear which
level (low complexity), M = 95.2 s, SD = 30.5 s (60.0– factor had the strongest influence on performance.
167.3 s) for the second level (medium complexity), and M = Consequently, we then calculated a multiple linear regression
116.5, SD = 33.6 s (62.5–199.6 s) for the last level (high com- (MLR) with average performance as the dependent variable
plexity). Consequently, and despite practice, the average du- and the user factors as independent variables (age, gender, GF,
ration of a level increased with complexity. SET, NfA) to identify the strongest predictors.
More interesting, the performance across the three levels in For the first game (physical exercise), the regression model
each game was strongly correlated and therefore consistent, explained 60% of the variance in performance (F(5,58) =
meaning that players were either slower or faster independent 18.763, p < .001, r2 = .603). For the second game (cognitive
of the level (α = .804 for the motion-based exercise game and training), the regression model accounted for 57% of the var-
α = .852 for the cognitive training game). These findings are iance in performance (F(5,39) = 9.913, p < .001, r2 = .573).
summarized in Table 2. Tables 4 and 5 show the regression models for the two games.
To evaluate how individual differences influenced perfor- For the motion-based game, the strongest predictor was age
mance, we first calculated an average performance score (β = −.598, p < .001**), with older people being slower than
across the three levels that would be used in the following. younger people by far, followed by participant gender (β =
As Table 3 shows, most user factors correlated with perfor- −.221, p = .013 < .05*) and NfA (β = .213, p = .022*), with
mance: In both games, performance was influenced by age women being slower than men and higher NfA yielding higher
(r = −.440 and r = −.543) and technical self-efficacy (r = .357 performance in the game. Neither SET nor participant GF had a
and r = .309), with age and lower self-efficacy yielding lower significant influence in the MLR model that accounted for the
performance. Gender, however, was not related to perfor- mutual influence among the predictor variables.
mance in either game. The reported GF was linked to perfor- For the cognitive training game, age was also the strongest
mance in the motion-based exercise game (r = .269), with predictor of performance (β = −.739, p < .001), followed by
Table 3 Correlation between user factors and attained performance in
Table 2 Descriptive statistics for the performance in the motion-based the motion-based exercise game (left) and the cognitive training game
exercise game (left) and the cognitive training game (right) (right) [gender dummy coded: 1 = male, 2 = female]
Level Exercise game [collected objects] Cognitive training game [s] Factor Average Average
performance game 1 performance game 2
Level 1 Level 2 Level 3 Low Medium High
Age −.440** −.543**
M 22.8 18.8 17.0 57.3 95.2 116.5 Gender
SD 6.5 4.8 4,9 18.7 30.5 33.6 SET .357** .309*
Min. 7 9 2 33.2 60.0 62.5 GF .269*
Max. 37 29 26 107.2 167.3 199.6 NfA .328** .318*J Public Health (Berl.): From Theory to Practice
Table 4 Regression table for performance in the motion-based exercise To understand what drives social acceptance of serious games
game (r2 = .573; gender dummy coded: 1 = male, 2 = female)
for healthcare in AAL environments, we first analyzed how in-
Variable B SE β t Value p Value dividual user factors, including attained performance in the game,
relate to the intention to use the game in the future. For both
(const) 24.825 3.120 – 7.958 .05). In the first game,
SET −.025 .459 −.007 −.055 .956 only GF influenced the intention to use (r = .445, p < .001), with
GF .887 .687 .141 1.292 .202 gamers reporting a higher intention to use the game. No other
NfA .990 .420 .213 2.357 .022* individual factors had a significant effect. For the second game,
none of the individual factors influenced the intention to use the
game, including no effect of GF (r = .005, p = .976 > .05). So,
although NfA had a positive effect on performance, no influence
GF (β = −.346, p = .047 < .05) and SET (β = .338, p = .020
between NfA and intention to use was found in either study.
< .05). Neither the participants’ gender nor their NfA influ-
Lastly, we studied how the evaluation dimensions from the
enced performance in the game.
UTAUT2 model related to the overall intention to use both
Qualitatively, we observed that, independent of their age,
games. In Fitness Farm, all evaluation dimensions had a signif-
players of both games enjoyed being able to control their own
icant influence on the intended use. In Cook It Right, all dimen-
speed without being incentivized or even punished: while
sions except effort expectancy were significantly related to the
some tried to be as fast as possible, others took their time.
intention to use. Since the criteria correlated strongly with each
Although both games provided feedback on the performance
other, we again identified the strongest predictors using an MLR.
(objects collected during a round or number of steps and time
For Fitness Farm, this procedure identified a model (F(2,61) =
taken), we did not show benchmarks for comparison (such as
16.090, p < .001) with two significant predictors and an overall
high scores). Here, user preferences should be implemented
explained variance in intention to use of slightly below 50%
individually: while some, mostly older adults, commented that
(r2 = .478). Specifically, habit (β = .503) and social influence
they liked not having to compare their performance against
(β = .301) were identified as the two strongest significant predic-
others, others found this particularly motivating.
tors. In Cook it Right, this procedure also identified a model with
In summary, the performance the participants attained was
two significant predictors (F(2,44) = 24.62, p < .001) and an ex-
mostly determined by their age, with older people being
plained variance of above 50% (r2 = .512). The two strongest
slower than younger people. This result may not come as a
predictors were hedonic motivation (β = .550) and, again, habit
surprise, so the important question of whether the perfor-
(β = .294). Figure 3 illustrates the correlations (r) and the stron-
mance achieved in the game influenced acceptance will be
gest predictors (β) from the MLR.
examined in the next section.
From a qualitative perspective, the participants had mixed
feelings about both games. For the physical exercise game, most
Usability and acceptance
participants liked the idea of embedding exercise in games, yet
some had doubts as to whether the current game was motivating
The overall intention to use the motion-based exercise game
in the long run. They suggested more activities within the game,
was high (M = 3.70/5, SD = 1.02) and above the center of the
the ability to connect with friends and to compete against each
scale (2.5). Likewise, the cognitive training game was also
other, and other game environments, such as walking through a
rated very positively (M = 3.40/5, SD = 1.11) and far above
city by exercising. While the comments on the cognitive exercise
the center of the scale.
game were more positive, a general concern was whether the
activities within the games would have actual benefits for health
Table 5 Regression table for performance in the game for cognitive and cognitive functioning. We suspect that there is a fine balance
training (r2 = .573; gender dummy coded: 1 = male, 2 = female) between a low usage barrier due to a simple and entertainment-
like user interface on the one hand and a credible medical effect
Variable B SE β t Value p Value
on the other.
(Const) 10.714 2.597 – 4.125J Public Health (Berl.): From Theory to Practice
Fig. 3 Significant correlations Fitness Farm Cook It Right
between the UTAUT2 predictor Social Influence
Physical Exercising Game Cognitive Training Game
variables and intention to use the
motion-based exercise game (left) 1 Facilitating
30 Conditions r=
and the cognitive training game =. .5
, 2 4
(right). Bold lines show the key 52 21 Price-Value r=. 0*
.5 r=. 33
3*
predictors and their strengths (β) r= Tradeo
1 r=.35
after a multiple linear regression r=.34 0*
Intention To Use Intention To Use
r=.653 =.503 Habit r=.540*, =.294
(r 2=.478) (r 2=.512)
r=.38 6*
1 r=.46 .550
r=. Performance =
43
8 Expectancy 2*,
= . 68
r= r
.2 Hedonic
32
Motivation
E ort
Expectancy
First, none of the participants in the two final evaluation Strikingly, intention to use was not influenced by either
studies had any serious difficulties in using the games and technical self-efficacy or NfA (but NfA influenced perfor-
successfully completing the three required levels. Although mance, so we measured correctly). Hence, the games were
we cannot formally prove this, we attribute this to the utiliza- perceived as worthwhile independent of the participant’s in-
tion of user-centered and participatory design principles, such terest in delivering high performance and independent of the
as personas guiding the development phases, early tests with participants ability to interact with complicated technical sys-
paper prototypes with different game concepts and interaction tems, such as AAL environments. Again, both may be attrib-
styles, and later frequent tests with functional prototypes. This uted to our design methodology. However, a different picture
is also reflected in the very positive overall evaluation and the might have emerged if we had not followed the various guide-
above-average intention to use the games. lines for designing games for the elderly. Games with compli-
Second, the results show that performance is explained by cated setup routines, complex menu structures, or overcrowd-
the age of the players, with older people being significantly ed interfaces may be a stronger barrier to performance, and
slower than younger players. This may not sound particularly here we expect an influence of, for example, technical self-
surprising, yet it shows that other frequently observed factors, efficacy on the intention to use.
such as lower technical self-efficacy and ability to interact Forth, in both games, the application of the UTAUT2
with information and communication technology, play a mi- showed that most factors of the model related to the intention
nor role in performance, but age-related differences in physi- to use the game again. Hence, the model is applicable, al-
ology and cognition. though other factors relevant for adopting medical technology
Third, the overall intention to use the games, based on the are missing (e.g., privacy disposition, trust in medical technol-
UTAUT2 technology acceptance model, is not directly related to ogy). Regarding the particular factors, habit was found to be
any of the investigated (explanatory) user factors aside from GF most influential in both games. Hence, people who envision
in Fitness Farm. On the one hand, this raises an important ques- integrating the game into their daily routine report higher in-
tion that is not sufficiently discussed in the serious game com- tention to use the game in the future. Consequently, the stron-
munity: How can we realize preventive healthcare interventions gest levers to increase the likelihood that the games are
for people not particularly inclined towards (computer) games? adopted are strategies to increase this perception. For Fitness
Can we work together with users to develop new forms of exer- Farm, social influence was also influential: people are more
cise activities that are not embedded within classic games, for likely to adopt the game if others (particularly peers) positive-
example, by embedding the suggested movements in other inter- ly appraise the game play. One way to realize this might be
actions, such as making music or art? What do we do when game communities in which family, friends, and others might
people have no interest in exercising at all—for example, do they connect and play together. In Cook It Right, hedonic motiva-
want to be nudged to be active more often? On the other hand, tion was the second strongest predictor for the acceptance of
this also shows that the attained performance—which was much the game: the wish to play the game again was directly linked
lower for older participants—is irrelevant for later acceptance of to the perceived fun, which some, but not all, participants
the games: as we did not give feedback on the player’s perfor- reported. Here, we suggest developing a variety of games so
mance compared to others (e.g., high score), slower players still that future inhabitants of AAL environments can freely choose
had fun with the game. the training games they like the most.J Public Health (Berl.): From Theory to Practice
Lastly, financing of (preventive) serious games for games. Thus, to address non-players, exercises also need to be
healthcare and the augmentation of living environments by integrated into other environments, and studies need to deter-
AAL technology need to be discussed. Currently, it is difficult mine whether a larger and more diverse range of measures has
to estimate the monetary benefits of such AAL environments a positive effect on social acceptance, use, and health and
and the long-term medical benefits of serious games for well-being.
healthcare. Consequently, more research is needed to
reduce the cost of these technologies and to make their
benefits quantifiable. Acknowledgements The authors thank all participants for their commit-
ment and valuable feedback. We thank our committed students Andreas
Schäfer, Jan Wittland, Dirk Nettelnbrecker, and Markus Gottfried for
their support during development and evaluation. We would like to thank
Limitations Kai Kasugai for taking the photos and Monica Gonzalez-Marquez for her
careful proofreading.
This work provides insights into the acceptance of future se-
Authorship contributions Both authors contributed to the design and
rious games for healthcare in AAL environments, yet it is not
implementation of the research, to the analysis of the results and to the
without its limitations. writing of the manuscript. Martina Ziefle supervised the project as a
First, the evaluation of the serious games for healthcare is whole, and Philipp Brauner coordinated the development of the serious
based on a short-term intervention. Hence, future work should games.
evaluate the long-term effectiveness of these interventions,
Funding Open Access funding enabled and organized by Projekt DEAL.
whether and how long positive effects will emerge and remain
This work was funded by the German Excellence Initiative.
after playing the game, whether and by whom the games will be
used over a longer period, and whether and to what extent health
Declarations
and overall well-being is improved. Prior work suggests that if
people use the games to exercise (see related work), positive Ethical approval and consent to participate Informed consent was ob-
effects should unfold. Also, our own work suggests that older tained from all individual participants included in the study. Participation
adults’ perceived pain decreases shortly after playing the physical was completely voluntary, was not rewarded, and was framed as usability
exercise game (Brauner and Ziefle 2020). On the other hand, the study, without suggesting any influence on the participants’ health.
anticipated AAL environments are not yet available, and only
Conflict of interest The authors declare that they have no conflict of
single AAL labs distributed across universities and other research interest.
institutions exist for user studies. Our approach of studying one
short-term interaction among many users enables deeper insights Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing, adap-
into usability and technology acceptance than studying few users tation, distribution and reproduction in any medium or format, as long as
over a longer period. you give appropriate credit to the original author(s) and the source, pro-
Second, to keep the length of the experiment reasonable, vide a link to the Creative Commons licence, and indicate if changes were
the set of (explanatory) user factors and evaluation dimensions made. The images or other third party material in this article are included
in the article's Creative Commons licence, unless indicated otherwise in a
was limited. Of course, there are numerous other user factors
credit line to the material. If material is not included in the article's
(e.g., disposition to trust, privacy disposition) and evaluation Creative Commons licence and your intended use is not permitted by
dimensions (e.g., perceived privacy of the system, trust in this statutory regulation or exceeds the permitted use, you will need to obtain
medical technology) that could influence social acceptance of permission directly from the copyright holder. To view a copy of this
licence, visit http://creativecommons.org/licenses/by/4.0/.
these serious games. Therefore, future studies should system-
atically identify relevant factors and weight their importance
in the respective factor space. This work, however, suggests
that the application of technology acceptance research models
is feasible for evaluating these systems, although additional
evaluation dimensions might be needed.
References
Third, in this article we refrained from detailed cor- Ahlskog JE, Geda YE, Graff-Radford NR, Petersen RC (2011) Physical
relation analyses between the (explanatory) user factors exercise as a preventive or disease-modifying treatment of dementia
and the evaluation dimensions. This is presented indi- and brain aging. Mayo Clin Proc Mayo Clin 86:876–884. https://
vidually for Fitness Farm (Brauner and Ziefle 2020) and doi.org/10.4065/mcp.2011.0252
Aldwin CM, Spiro A, Park CL (2006) Health, behavior, and optimal
Cook It Right (Wittland et al. 2015).
aging: a life span developmental perspective. In: Birren JE, Schaie
Lastly, this work presented a game-based approach for ex- KW (eds) Handbook of the psychology of aging. Elsevier Academic
ercise that embedded physical and cognitive exercises within Press, Burlington, pp 85–104
two different game environments. Yet, the findings show that Arning K, Ziefle M (2007) Understanding age differences in PDA accep-
non-players are not easily motivated to exercise by means of tance and performance. ComputHuman Behavior 23:2904–2927J Public Health (Berl.): From Theory to Practice
Ball KK, Wadley VG, Vance DE, Edwards JD (2004) Cognitive Skills: meta-analysis. Archives Gen Psychiat 55:809–815. https://doi.org/
Training, Maintenance, and Daily Usage. Encyclopedia Appl 10.1001/archpsyc.55.9.809
Psychol 387–392 Gentry SV, Gauthier A, Ehrstrom BLE, et al (2019) Serious Gaming and
Beul S, Ziefle M, Jakobs E-M (2012) How to bring your doctor home. Gamification Education in Health Professions: Systematic Review.
Designing a telemedical consultation service in an ambient assisted J Med Internet Res
living environment. In: Duffy VG (ed) Advances in human aspects Gerling K, Livingston I, Nacke L, Mandryk R (2012) Full-body motion-
of healthcare. CRC Press, Boca Raton based game interaction for older adults. In: Proceedings of the 2012
Bonnechère B, Jansen B, Omelina L, Van Sint Jan S (2016) The Use of ACM annual conference on human factors in computing systems -
Commercial Video Games in Rehabilitation: A Systematic Review. CHI ‘12. ACM Press, New York, pp 1873–1882
Int J Rehab Res Giannakouris K (2008) Ageing characterises the demographic perspec-
Brauner P, Ziefle M (2020) Serious motion-based exercise games for tives of the European societies
older adults: evaluation of usability, performance, and pain mitiga- Guiney H, Machado L (2013) Benefits of regular aerobic exercise for
tion. JMIR Serious Games 8:e14182. https://doi.org/10.2196/14182 executive functioning in healthy populations. Psychon Bull Rev
Brauner P, Holzinger A, Ziefle M (2015) Ubiquitous computing at its 20:73–86. https://doi.org/10.3758/s13423-012-0345-4
best: serious exercise games for older adults in ambient assisted Hader C, Beischer W, Braun A et al (2004) Diagnostik, Therapie und
living environments – a technology acceptance perspective. EAI Verlaufskontrolle des Diabetes mellitus im Alter. Diabetes und
Endorsed Trans Serious Games 15:1–12. https://doi.org/10.4108/ Stoffwechsel 13:31–56
sg.1.4.e3 Heyn P, Abreu BC, Ottenbacher KJ (2004) The effects of exercise train-
Choi E, Shin SH, Ryu JK, et al (2020) Commercial video games and ing on elderly persons with cognitive impairment and dementia: a
cognitive functions: Video game genres and modulating factors of meta-analysis. Archives Phys Med Rehab 85:1694–1704 https://doi.
cognitive enhancement. Behavior Brain Functions 16. https://doi. org/10.1016/j.apmr.2004.03.019
org/10.1186/s12993-020-0165-z Ho KKL, Pinsky JL, Kannel WB, Levy D (1993) The epidemiology of
Cook DJ, Das SK (2007) How smart are our environments? An updated heart failure: the Framingham study. J Am College Cardiol 22:A6–
look at the state of the art. Pervasive Mobile Comput 3:53–73. A13
https://doi.org/10.1016/j.pmcj.2006.12.001
Ijsselsteijn W, Nap HH, de Kort Y, Poels K (2007) Digital Game Design
Cook DJ, Augusto JC, Jakkula VR (2009) Ambient intelligence: technol- for Elderly Users. In: Future play ‘07. ACM, Toronto, pp. 17–22
ogies, applications, and opportunities. Pervasive Mobile Comput 5:
Kasugai K, Ziefle M, Röcker C, Russel P (2010) Creating spatio-
277–298. https://doi.org/10.1016/j.pmcj.2009.04.001
temporal contiguities between real and virtual rooms in an assistive
Cooney GM, Dwan K, Greig CA, et al (2013) Exercise for Depression.
living environment. In: Bonner J, Smyth S, O’ Neill S, Mival O
Cochrane Database Syst Rev 9. https://doi.org/10.1002/14651858.
(eds) Proceedings of Create 10 Innovative interactions.
CD004366.pub6
Loughborough, Elms Court., pp. 62–67
Czaja SJ, Charness N, Fisk AD et al (2006) Factors predicting the use of
Klack L, Möllering C, Ziefle M, Schmitz-Rode T (2011a) Future care
technology: findings from the Center for Research and Education on
floor: a sensitive floor for movement monitoring and fall detection in
aging and technology enhancement. Psychol Aging 21:333–352^.
home environments. Wireless Mobile communication and
https://doi.org/10.1037/0882-7974.21.2.333
healthcare, lecture notes of the Institute for Computer Sciences, so-
Davis FD (1993) User acceptance of information technology: system cial informatics and telecommunications. Engineering 55:211–218
characteristics, user perceptions and behavioral impacts. Int J Man-
Klack L, Schmitz-Rode T, Wilkowska W et al (2011b) Integrated home
Machine Stud 38:475–487. https://doi.org/10.1006/imms.1993.
monitoring and compliance optimization for patients with mechan-
1022
ical circulatory support devices. Ann Biomed Eng 39:2911–2921
Deterding S, Dixon D, Khaled R, Nacke L (2011) From Game Design
Kleinberger T, Becker M, Ras E et al (2007) Ambient intelligence in
Elements to Gamefulness: Defining Gamification. In: MindTrek ‘11
assisted living: enable elderly people to handle future interfaces.
Proceedings of the 15th International Academic MindTrek
Universal Access in HCI, In, pp 103–112
Conference Envisioning Future Media Environments. pp 9–15
Elliot AJ, Dweck CS (eds) (2005) Handbook of competence and motiva- Knight JA (2012) Physical inactivity: associated diseases and disorders.
tion. The Guilford Press, New York Annals of Clinical & Laboratory Science 42:320–337
Elwood P, Galante J, Pickering J, et al (2013) Healthy lifestyles reduce Konstantinidis EI, Billis AS, Mouzakidis CA et al (2016) Design, imple-
the incidence of chronic diseases and dementia: Evidence from the mentation, and wide pilot deployment of FitForAll: an easy to use
Caerphilly cohort study. PLoS ONE 8. https://doi.org/10.1371/ exergaming platform improving physical fitness and life quality of
journal.pone.0081877 senior citizens. IEEE Journal of Biomedical and Health Informatics
Fisk AD, Rogers WA, Charness N et al (2009) Designing for older adults: 20:189–200 https://doi.org/10.1109/JBHI.2014.2378814
principles and creative human factors approaches. CRC Press, Boca Lampit A, Hallock H, Valenzuela M (2014) Computerized cognitive
Raton training in cognitively healthy older adults: a systematic review
Flandorfer P (2012) Population ageing and socially assistive robots for and meta-analysis of effect modifiers. PLoS Med. https://doi.org/
elderly persons: the importance of sociodemographic factors for user 10.1371/journal.pmed.1001756
acceptance. Int J Population Res. https://doi.org/10.1155/2012/ Leusmann P, Möllering C, Klack L et al (2011) Your floor knows where
829835 you are: sensing and Acquisition of Movement Data. IEEE
Fleming TM, Bavin L, Stasiak K et al (2017) Serious games and International Conference on Mobile Data Management, In, pp 61–
gamification for mental health: current status and promising direc- 66
tions. Frontiers Psychiat. https://doi.org/10.3389/fpsyt.2016.00215 Lindley SE, Harper R, Sellen A (2008) Designing for elders: exploring
Franco JR, Jacobs K, Inzerillo C, Kluzik J (2012) The effect of the the complexity of relationships in later life. In: Proceedings of the
Nintendo Wii fit and exercise in improving balance and quality of 22nd annual conference of the British HCI group (HCI 2008), pp
life in community dwelling elders. Technol Health Care 20:95–115. 77–86
https://doi.org/10.3233/THC-2011-0661 López-Otín C, Blasco M a., Partridge L, et al (2013) The hallmarks of
Gao S, Hendrie HC, Hall KS, Hui S (1998) The relationships between aging. Cell 153:1194–1217. https://doi.org/10.1016/j.cell.2013.05.
age, sex, and the incidence of dementia and Alzheimer disease: a 039You can also read
